Hi,

I am trying to setup a log shipping replication on a system and i was
wondering if the documentation was up to date. Please let me know if
i miss something of theses steps :

- I need to replicate from Master to SLAVE, slave is the log shipping
  destination.

- To do so i have to setup a real node and i am doing it on the same
  postgresql DB, i am calling it MIRROR.

- I synchronize the schemas from MASTER to SLAVE and MIRROR.

- I setup two slon process one for the SLAVE and one for the MIRROR

- I setup -a on the slon process of the MIRROR to dump the log files.

- I create the sets and subscribe to it (from MASTER TO MIRROR).

- I launch the slony_dump1.sh and dump the sql into the SLAVE.

- I synchronize all theses dump files from the dump directory to slaves
  every few minutes or so.

Everything looks working well my only concerns is the part of the
documentation advising to analyze the first few headers and see if it
is commiting before commiting the rest of the files.

I had no problems to just :

for i in *;do psql -hSLAVEDB -f $i db && rm -f $i;done

without needing to analyze the headers.

So is it still needed to analyze the headers or something got fix in
the latest version of slony 1.2.10

Cheers, Chmouel.
-- 
http://www.squiz.net